The topic of nova explosions introduces areas of physics not touched upon in previous chapters: white dwarf structure, nuclear reactions, hydrodynamics of explosive mass loss, common envelope structures, non-LTE conditions in low density ejecta, dust formation. At the same time these are areas that have been comprehensively reviewed and referenced in recent books and articles. This chapter therefore concentrates on the basics of eruption physics, expanding only in those parts that relate strongly to the properties and evolution of the classes of CVs discussed in other chapters. Novae between eruptions are discussed in Chapter 4.

Two books provide detailed reviews of observation and theory of nova eruptions: The Galactic Novae (Payne-Gaposchkin 1957) and Classical Novae (Bode & Evans 1989). The latter contains a comprehensive list of references to observational papers on novae published to the beginning of 1987. Modern review articles, and a few early ones still of value, are Stratton (1928), McLaughlin (1960a), Gallagher & Starrfield (1978), Truran (1982), Bode & Evans (1983), Starrfield (1986, 1988, 1990, 1992), Starrfield & Snijders (1987), Gehrz (1988), Shara (1989), Seitter (1990). Specialized conference proceedings appear in Friedjung (1977) and Cassatella & Viotti (1990).

Lists of novae, finding charts and references are given in Duerbeck (1987), Bode & Evans (1989) and Downes & Shara (1993).

Nova Discovery

Prior to the introduction of wide-field sky photography in the late nineteenth century novae were found generally as naked eye objects.

A variety of stars was early recognised to have eruptions that bear some resemblance to those of novae. Consequently, a very heterogeneous class of nova-like variables (NLs) was introduced (see, e.g., Campbell & Jacchia 1941; Kukarkin et al. 1958; Petit 1987), which, with hindsight, is seen to include many types of object that are totally unrelated structurally to the true novae. For example, η Car, γ Cas and P Cyg stars do not have the duplicity of CVs. Symbiotic stars, on the other hand, may all be close binaries and some may contain degenerate components as in the CVs (Kenyon 1986).

With the removal of these eruptive objects to their own classes, paradoxically only the non-eruptive residue remained as NLs (many of which, however, may have ‘low states’). It is from their short time scale spectroscopic and photometric behaviour (including the evidence of binary structure) that such stars are recognized as resembling novae between eruptions (they could more appropriately have been termed ‘NR-like’).

Classifications

From the incomplete discovery of novae in earlier centuries (Section 1.1) it is clear that among the NLs there should be many unrecognized NRs. Similarly, the fact that many novae discovered this century can be found as blue objects on archival sky survey plates shows that among the currently known NLs must be a number of pre-novae.

This first chapter is designed to give the reader an historical perspective on the subject of cataclysmic variable (CV) stars. Ground-based photometric and spectroscopic observational developments up to 1975 are treated in detail. Since that date instrumental methods in the optical region have been to some extent fixed, and to continue the historical approach would be repetitive of much of what appears in later chapters. The introduction of observational techniques in other wavelength regions is, however, followed beyond 1975

Pre-1900 Observations of Novae

If the ancient philosophers had been correct in their assertion that the distant stars are immutable, incorruptible and eternal, astronomy would be the dullest of disciplines. Fortunately, they were wrong on all counts. The stars possess variability on all time scales and amplitudes, sufficient to satisfy all interests, from the exotic to the commonplace, from the plodding to the impatient.

Among these, the most prominent celestial discordants are the novae Stella: new stars, challenging the ancients in their own times, but, such was the power of Aristotelian philosophy, passing almost entirely unacknowledged in European and Middle Eastern societies until the post-Copernican era (Clark & Stephenson 1977). In China, however, records of celestial events (kept mostly for astrological purposes) have been maintained since c. 1500 BC, and there are supporting and supplementary records in Japan from the seventh century AD and in Korea from c. 1000 AD (Clark & Stephenson 1976, 1977). Among these are numerous accounts of temporary objects, from which may be sifted comets, meteors, novae and supernovae.

The history of cataclysmic variable star research mirrors the objects themselves: periods of relative inactivity punctuated by heightened or even explosive advances. Until about 1970 each resurgence of interest was a result of a distinct technological advance. In the past two decades the technological improvements have been almost continuous and the interest in cataclysmic variables has burgeoned from the realization that they have so much to offer. Not only are they of interest per se, exhibiting a challenging range of exotic phenomena covering the electromagnetic spectrum from radio waves to TeV gamma rays, and time scales from fractions of a second to millions of years, they are important for their relevance to other exciting areas of astrophysics.

For example, it has become evident that accretion discs are one of the most commonly occurring structures – probably all stars form from disc-like configurations, with material left over to provide planetary systems. A large fraction of binary stars form accretion discs at some stage of their evolution. Accretion discs are important in X-ray binaries – matter accreting onto neutron stars or black holes. Entire galaxies are initially gaseous discs, and most may develop central discs intermittently that fuel their active nuclei.

But it is in cataclysmic variables (CVs) that accretion discs are observed to best advantage – quasi-stable discs, unstable discs and transformations between them. In dwarf novae during outburst, or in nova-like variables in their high state, the light is dominated by emission from discs – and being almost two-dimensional their observed properties are strongly affected by the viewing angle. All are close double stars, and those with eclipses present unrivalled opportunities for determining spatially resolved physical structures.

From systems that are weakly or covertly magnetic we turn to ones in which the magnetic field of the primary is strong enough to control the accretion flow, preventing the formation of an accretion disc and generating the signatures of magnetic accretion: large linear and circular optical polarization and strong X-ray emission.

Historical Development

The discovery of the polars provides a lesson that even relatively familiar objects may reveal exotic phenomena if interrogated in the correct way. The star AM Her had been discovered as a variable in 1924 and listed as a NL on the basis of slow variations in brightness over a range of 3 mag and an emission-line spectrum. In 1976 Berg & Duthie (1977) suggested that AM Her could be the optical counterpart of the Uhuru X-ray source 3U 1809+50 and Hearn, Richardson & Clark (1976) using the SAS-3 satellite found a variable soft X-ray source near the same position. The similarity of this source to the low mass X-ray binaries Sco X-l and Cyg X-2 stimulated Cowley & Crampton (1977) to obtain spectra, which revealed a 3.09 h orbital period.

The main surprise came, however, when Tapia discovered in August 1976 that AM Her is linearly and circularly polarized at optical wavelengths (Tapia 1977a). Its linear polarization varies from zero up to 7% and its circular polarization from −9% to +3%, both changing smoothly over the period of 3.09 h (Figure 1.12). The high degree of circular polarization, previously only seen in magnetic white dwarfs (Angel 1978), suggested the presence of a strong magnetic field.

The DN, already introduced in Sections 1.2 and 1.3 with their classification scheme described in Section 2.1, are arguably the most valuable of objects for the study of accretion discs. Among them examples may be found of optically thin discs and optically thick discs, of face-on discs and edge-on discs, of non-steady discs and of nearly steady state discs and of transitions between them. Furthermore, the brightest DN at maxima reach apparent magnitudes of 8–10, at which time the entire flux is conveniently of almost pure accretion origin.

Well-Observed DN

It is inevitable that a few relatively bright DN, especially the eclipsing systems, have been preferentially observed. Although over 200 DN have been classified by their light curves, only a small fraction have been studied sufficiently to establish their orbital periods. It will be seen in this chapter that POrb plays an important rôle in the systematics of DN. Among the DN in general, 12 Z Cam stars, 29 definite U Gem stars (including, slightly unconventionally, the three systems BV Cen, GK Per and V1017 Sgr with large POrb), 34 SU UMa stars and 22 objects suspected of belonging in the DN class have known orbital periods. The SU UMa stars may be overrepresented because their orbital periods are easy to estimate, independent of inclination, from photometric observations made during super outbursts. Orbital periods for the U Gem and Z Cam class have come predominantly from spectroscopic observations, with the addition of a few found from photometric orbital variations (eclipses, bright spot modulation, IR ellipsoidal modulation).

The many measurements of the spectrum of the CBR discussed in the preceding chapter are consistent with a Planck spectrum with T0 = 2.73 ± 0.02 K over a wavelength range 0.1 cm ≲ λ≲75 cm. Only the submillimeter observations reviewed in Section 4.8.4 provided any evidence for a significant deviation from a thermal spectrum, and these appear to be erroneous. What conclusions may we draw from the essentially thermal spectrum of the CBR? This chapter provides some answers to that question. It is presented as an introduction to, not an exhaustive treatment of, the processes that determine the CBR spectrum. There are a number of reviews, which treat these topics in more detail, such as Danese and De Zotti (1977), Sunyaev and Zel'dovich (1980) and Bond (1988).

We begin by noting the conditions under which we would expect an exactly thermal spectrum to have been produced early in the Hot Big Bang, then consider a number of physical processes that could have distorted an initially thermal spectrum. We also consider the possibility that one or more additional ‘cosmic’ backgrounds may be present, adding to the CBR at wavelengths below about 1 mm. Finally, we investigate the constraints that the spectral measurements of Chapter 4 place on these processes.

A characteristic feature of the CBR, noted at the time of its discovery by Penzias and Wilson (1965), is its approximate isotropy (see Appendix A). Approximately equal intensity in all directions is expected if the radiation is a relic of the Hot Big Bang. On the other hand, there are a variety of mechanisms that can induce small amplitude variations in intensity, or anisotropies, into an initially uniform CBR; some of these were outlined in Chapter 2 and will be discussed in detail in Chapter 8.

Careful measurements of the angular distribution of the CBR have therefore been pursued both to confirm the cosmic, Hot Big Bang, origin of the CBR and to search for small amplitude anisotropies imprinted in it. In this chapter we deal with observations of the angular distribution of the CBR on the largest angular scales, θ ≳ 10°, and in particular with the dipole and quadrupole moments of the CBR. The value of about 10° for the boundary between ‘large’ scale anisotropies (discussed in this chapter) and smaller scale anisotropies (Chapter 7) is obviously rather artificial. When we turn in Chapter 8 to the implications of the measurements of and upper limits on CBR anisotropies, we will be drawing on the results of both Chapters 6 and 7.

One of the basic problems of cosmology is the singularity characteristic of the familiar cosmological solutions of Einstein's field equations. Also puzzling is the presence of matter in excess over antimatter in the universe, for baryons and leptons are thought to be conserved. Thus, in the framework of conventional theory we cannot understand the origin of matter or of the universe. We can distinguish three main attempts to deal with these problems.

1. The assumption of continuous creation (Bondi and Gold 1948; Hoyle 1948), which avoids the singularity by postulating a universe expanding for all time and a continuous but slow creation of new matter in the universe.

2. The assumption (Wheeler 1964) that the creation of new matter is intimately related to the existence of the singularity, and that the resolution of both paradoxes may be found in a proper quantum mechanical treatment of Einstein's field equations.

3. The assumption that the singularity results from a mathematical over-idealization, the requirement of strict isotropy or uniformity, and that it would not occur in the real world (Wheeler 1958; Lifshitz and Khalatnikov 1963).

If this third premise is accepted tentatively as a working hypothesis, it carries with it a possible resolution of the second paradox, for the matter we see about us now may represent the same baryon content of the previous expansion of a closed universe, oscillating for all time.

The science that treats the properties and evolution of the Universe as a whole is cosmology. Among the sciences, it is unique in having only a single object of study – there are no other Universes for us to use as controls, nor can we readily run the whole experiment over again. As a consequence, much of the effort in modern cosmology has been to determine the best mathematical description, or ‘model’, of the Universe we inhabit. As we shall see, that task is not yet complete, despite the rapid advances of the past few decades. The range of possible models is presented later in this chapter. First, though, we need to look at the observational bases of modern cosmology, a set of astronomical observations which have established the Hot Big Bang theory and restricted the range of models we need to consider.

Astronomical constituents of the Universe

Since cosmology is the study of the Universe as a whole and as a single system, it is only indirectly concerned with subsystems within the Universe. Here, I will mention only two: galaxies and clusters of galaxies. The galaxies are assemblies of 108–1012 stars; many galaxies also contain appreciable amounts of interstellar gas and dust.

The very first astronomical signal at radio wavelengths, by happy coincidence, was also detected at the Bell Telephone Laboratories; in 1932, Karl Jansky detected at 15 m wavelength radio emission, which he correctly identified as coming from the Galactic plane. Astronomers paid little attention. Observational radio astronomy did not really come into its own until after World War II (see, e.g., Hey, 1973, and Sullivan, 1984). It is now recognized as a powerful adjunct to optical astronomy, particularly in the study of low density cosmic matter and of energetic objects and phenomena such as quasars and the collimated jets seen in radio galaxies. We will look very briefly at radio sources later in this chapter, but most of it will be devoted to the tools and techniques of observational radio astronomy. Chapter 3 is designed to introduce the more specialized radio astronomical techniques used in studying the CBR; it is not intended to be a complete introduction to radio astronomy. For further details, readers may want to consult one or more of the following texts: Kraus (1986); Rohlfs (1986); and Christiansen and Högbom (1985). Interferometry is very fully treated by Thompson, Moran and Swenson (1986), and radio sources by Pacholczyk (1970) and Verschuur and Kellermann (1988), among others. The treatment of radio astronomy in this book is closest to the work of Rohlfs.

Humankind has made stories about the origin of the world since prehistoric times. These creation stories often have a grand beauty and are sometimes richly detailed. It is only in the present century that such myths and images have been supplanted by a well-established scientific description of the origin of the world. ‘World’ is now understood to mean the Universe as a whole, not just the Earth or the solar system, and the modern picture of its origin and evolution is the Hot Big Bang model. This book describes one crucial piece of astronomical evidence supporting the Big Bang model, namely the cosmic microwave background radiation, heat radiation left over from a hot and dense phase early in the history of the Universe.

The cosmic background radiation (CBR) was discovered, by accident as it happens, a quarter of a century ago. Within a few years, the basic properties of the radiation had been established. Those properties, especially the thermal 3 K spectrum and the very uniform distribution of the CBR across the sky, have convinced virtually all astrophysicists that the radiation is a relic of the Hot Big Bang, and that it comes to us from a very early time in the history of the Universe. It thus provides information about the early history of the Universe obtainable in no other way.

The publishers have kindly agreed to allow me to include a brief appendix in order to update some of the observational results contained in Chapters 4, 6 and 7. As noted in the preface, references in the main body of the text are complete to early 1992; here I provide summaries of some of the results reported in 1992, 1993 and early 1994. In the space of a few pages, I cannot hope to cover all the work in this field, in which hundreds of papers were published in that time interval. Instead, I will emphasize the most crucial observational results, and provide a few references to theoretical papers which establish new results or provide good reviews of the field.

Spectrum

The COBE team (Mather et al., Fixsen et al., and Wright et al., all 1994) have reported new and more complete results on the short-wavelength spectrum of the CBR. From a direct measurement of the spectrum using the FIRAS instrument (Section 4.9), Mather et al. (1994) determine T0 = 2.726±0.010 K. Their observations allow them to place limits on distortions of the spectrum; in the notation introduced in Chapter 5, these limits are |y| <2.5×10-5 and |µ| <3.3×10-4. These values, like the error in T0, are given at the 95% confidence level.

Soon after the discovery of the CBR it was recognized that measurements of, or upper limits on, its anisotropy on scales of degrees or less would provide unique information about the origin and development of structure within the Universe. Such observations are of particular value because they probe cosmic times well before the appearance of any luminous objects in the Universe such as galaxies or stars. The surface of last scattering from which the CBR photons reach us is more distant than any QSO or galaxy yet detected (these lie at z < 5), and the CBR thus carries information about the state of the Universe at earlier times. It is quite likely, in fact, that the CBR photons we study reach us from the first few hundred thousand years of the history of the Universe; and, as we shall see in Chapter 8, the CBR may encode information from the epoch of inflation nearly 50 orders of magnitude earlier still.

With so much to learn, astronomers have worked hard to detect fluctuations in the temperature of the CBR. Indeed, there have been as many searches for structure in the background on scales ≤ 10° as measurements of the dipole and of the spectrum combined.